CFAES Give Today

Department of Agricultural, Environmental, and Development Economics


National Perspectives-Maps


Method for creating the above maps:

We begin by defining the outer boundary of the exurban field using the US Bureau of the Census 2003 defined Metropolitan Statistical Areas (MSA). There are 356 MSAs in the lower 48 states, comprised of 1,080 counties. MSAs are geographic areas that consist of the county or counties associated with at least one core urbanized area with a population of at least 50,000, plus adjacent counties having a high degree of social and economic integration with the core, as measured through commuting ties with the counties containing the core. Thus, the MSA delineation provides an approximate geographical extent of the commutershed that corresponds to large urbanized areas in the U.S.

To isolate the exurban field within MSAs—the regions of the US in which exurban development takes place, we omit the more densely populated areas. In the US, urbanized areas are densely settled areas (typically at least 1,000 people per square mile) with a population of at least 50,000 people. Using Geographic Information Systems (GIS), we remove the urbanized areas from the MSAs. In addition, we omit non-developable land by removing major water bodies and federal lands (including National Forests, Bureau of Land Management lands, National Wildlife Refuges, National Parks and Wilderness Areas). The remainder area is the less densely populated area of MSAs, what we refer to as the U.S. exurban field. This method of identifying the exurban field satisfies three primary objectives: first, it delineates the exurban field based on the notion of a potential commutershed; second, it captures those areas that that are dependent on urbanized areas, but are not already urbanized; and third, it overcomes the problem of under-bounding and over-bounding discussed earlier.

With this definition in hand, data from the 2003 LandScan population distribution model, created by the US Department of Energy’s Oak Ridge National Laboratory (UT Battelle, LLC., 2005), are used to describe the spatial characteristics of exurban settlement patterns located within the exurban field. The LandScan model estimates worldwide ambient populations at a 30” by 30” resolution (approximately 0.69 square km in the lower 48 states), which is the finest-scale global population data produced to date (Bhaduri, Bright, Coleman and Dobson, 2002). The model spatially allocates population on this grid by assigning a probability coefficient to each cell which is then applied to census counts. The probability coefficients for each cell are based on factors that contribute to population density, e.g. transportation networks, land cover, slope, and nighttime lights. This dataset has been described as making a “foundational” contribution to future social economic and demographic study (Sutton, Elvidge, Obremski, 2003). Nonetheless, it should be noted that these are population density estimates and not actual counts. Despite this weakness, the LandScan data provide a consistent, fine-scale representation of population density on a regular grid for the entire U.S. For this reason, we conclude that the advantage of using the LandScan data for national comparison of regional exurban patterns far outweighs the limitation.

To identify the pattern of exurban settlement using data on population density, a classification scheme is necessary. We assign a density class and settlement type for each cell according to the following categorization:

Density Class Settlement Type People/sq mi acres/household*
Very low Rural/Wilderness 0-10 165 or greater
Low Rural 10-100 16.5-165
Medium Exurban 100-1,000 1.6-16.5
High Suburban/urban 1,000+ less than 1.6

* Based on average household size in the U.S.

While any classification scheme is to some extent ad hoc, the definition for the medium density class used here to quantify exurban settlement patterns generally reflects lot sizes typified by the “hobby farms” that can support houses without sewer connections, in addition to reflecting densities suggested by other researchers (Daniels, 1999; Theobald, 2001; Wolman et al., 2005).


Bhaduri, B.L., E.A. Bright, P.R. Coleman, and J.E. Dobson. (2002). LandScan: Locating.
people is what matters. GeoInformatics 5(2), 34-37.

Daniels, T.L. (1999), When city and country collide: Managing growth in the metropolitan fringe. Washington, DC: Island Press.

Sutton P.C., C. Elvidge, and T. Obremski (2003). Building and evaluating models to estimate ambient population density. Photogrammetric Engineering and Remote Sensing 69(5), 545-553.

Theobald, D.M. (2001). Land-use dynamics beyond the American urban fringe. Geographical Review 91(3), 544-565.